Document 12784985

advertisement
Plasticity
http://imechanica.org/node/17162
Z. Suo
RHEOLOGICAL MODELS
Rheology is the science of deformation. This science poses a question for
every material: Given a history of stress, how do we predict the history of strain,
or the other way around?
Rheology answers this question using mathematical models. In
constructing a rheological model, we let material evolve through a sequence of
homogeneous, but multiaxial, states of stress and strain. We represent a history
of stress by a tensor as a function of time, and represent a history of strain by
another tensor as a function of time.
For a given material, we fit a rheological model to the experimental data
of some histories of stress and strain, and then use the model to predict other
histories of stress and strain.
Deformation is often large, and rheological models are usually nonlinear.
In this course we have studied several most successful rheological models: linear
viscosity, nonlinear viscosity, viscoplasticity, and perfect plasticity. Here we
summarize these models, and then construct several others: strain hardening,
elastoplasticity, and viscoelasticity. In a separate course, we study largedeformation elasticity, electroelasticity, and poroelasticity (Suo 2013).
We will focus on models in their simplest forms. For deeper and more
extensive discussions, see review articles and textbooks (e.g., Tanner 2000;
Nemat-Nasser 2004; Xiao, Bruhns, Meyers 2006; Gurtin, Fried, Anand 2010;
Irgens 2014).
The Second-Invariant Model
Incompressibility. When a material undergoes a large deformation,
the constituents of the material—atoms, molecules and colloidal particles—
change neighbors. The large deformation causes large change in shape, but
nearly no change in volume. Unless otherwise stated, we neglect the small
change in volume, and adopt the idealization that the material is incompressible.
Let D be a rate of deformation. The condition of incompressibility is Dkk = 0 .
Hydrostatic stress does not affect deformation. Most models
adopt a further idealization: Superposing a hydrostatic stress on a material does
not affect its deformation. For example, such models neglect the effect of
pressure on viscosity and yield strength. Let σ ij be a state of stress. The mean
stress is σ m = σ kk / 3 , and the deviatoric stress is sij = σ ij − σ mδij . Unless otherwise
stated, we assume that deformation is entirely caused by the deviatoric stress,
and is independent of the mean stress.
December 7, 2014
rheological models 1
Plasticity
http://imechanica.org/node/17162
Z. Suo
Isotropy. We will adopt yet another idealization: the material is
isotropic. We formulate models in terms of the second invariant of the deviatoric
stress, sij sij / 2 . We neglect the effect of the third invariant.
In a shearing flow, of the rate of shear γ and the shearing stress τ , the
tensors of the rate of deformation and the deviatoric stress take the forms:
! 0 γ / 2 0 $
! 0 τ 0 $
#
&
#
&
s =# τ 0 0 &.
D = # γ / 2 0 0 & ,
#
&
# 0 0 0 &
0 0 &
#" 0
"
%
%
In the shearing flow, the second invariant of the deviatoric stress is sij sij / 2 = τ 2 .
This coincidence motivates the following definition. For a flow under a
general state of deviatoric stress s, define the equivalent shearing stress by
τ e = sij sij / 2 .
The equivalent shearing stress τ e is just another way to write the second
invariant of the deviatoric stress, and is a scalar measure of the amplitude of the
state of stress.
The second-invariant model. Here we focus on the model of the form
D = λs , where λ is a scalar. For a linearly viscous material, λ relates to the
viscosity by λ = 1 / 2η . For other types of rheological behavior, including
nonlinear viscosity and plasticity, λ depend on the state of stress. The secondinvariant model assumes that λ depend on the state of stress through the second
invariant, λ τ e . We fit this function to the experimental data measured in a
( )
simple type of stress, such as shear, or uniaxial tension.
The deviatoric stress is traceless, skk = 0 , and ensures that the material is
incompressible, Dkk = 0 . Furthermore, the model requires that λ ≥ 0 , and thus
satisfies the thermodynamic inequality sij Dij ≥ 0 for any state of stress.
The equation D = λs dates back at least to the theory of linear viscosity
(Stokes 1845). In describing other types of rheological behavior, this model is
known as the J2 flow theory, the Levy-Mises flow rule, or the generalized
Newtonian fluid.
We have seen this model as a special case of Rayleigh’s dissipation
functions. We have also linked this model to the mathematics of convex
functions (Suo 2014, notes on nonlinear viscosity).
Viscous Flow
December 7, 2014
rheological models 2
Plasticity
http://imechanica.org/node/17162
Z. Suo
Viscous flow. The model of viscosity answers the fundamental question
of rheology in the purest form: The rate of deformation is a one-to-one function
of the deviatoric stress. When the stress vanishes, the flow stops instantly, with
no delay. When the stress changes, the rate of shear changes instantly, with no
delay.
We test a material under a simple state of stress, such as shear. We plot
the experimental data as the flow curve, the relation between the shearing stress,
τ , and the rate of shear, γ . The flow curve corresponds to a monotonic function,
γ = h τ .
()
τ
viscous
γ
The second-invariant model relates two tensors, D = λs . For a linearly
viscous material, λ is a material constant, and relates to the viscosity as
λ = 1 / 2η . For a nonlinearly viscous material, λ is stress-dependent and is a
( )
function of the equivalent shearing stress, λ τ e .
( )
For a given material, we fit the function λ τ e
to the flow curve
determined experimentally under one type of stress, such as shearing stress.
When the fluid is under any state of stress, the second-invariant model predicts
that the rate of deformation relates to the deviatoric stress as
h τe
D=
s.
2τ e
( )
This general model recovers the flow curve γ = h τ
()
determined under the
shearing stress. So long as the flow curve γ = h τ is a monotonically increasing
()
function, the second invariant model specifies a one-to-one relation between the
two tensors: the rate of deformation D and the deviatoric stress s
Viscoplastic flow. The model of viscosity is applicable even when the
flow curve γ = h τ is not a one-to-one relation between the shearing stress and
()
the rate of shear. A viscoplastic fluid has a particular type of flow curve, with a
yield stress k. The flow curve is not a one-to-one relation between the rate of
December 7, 2014
rheological models 3
Plasticity
http://imechanica.org/node/17162
Z. Suo
shear and the shearing stress: A range of stress τ < k gives a single value of the
rate of shear, γ = 0 . When the stress exceeds the yield stress, τ > k , the shearing
stress and the rate of shear has a one-to-one relation.
()
Thus, h τ = 0 when
()
τ < k , and h τ is a monotonic function when τ > k .
τ
+k
γ
−k
In a multiaxial state of stress, the second invariant model relates the
tensor of rate of deformation to the tensor of deviatoric stress:
h τe
D=
s.
2τ e
( )
This model of viscoplasticity looks identical to that of nonlinear viscosity. For a
viscoplastic material, however, the model no longer prescribes a one-to-one
relation between the state of rate of deformation and the state of deviatoric stress.
Given a state of stress s, we calculate the equivalent shearing stress,
τ e = sij sij / 2 . When the state of stress is inside the yield surface, τ e < k , the
material is rigid, and the rate of deformation vanishes, D = 0 . When the state of
stress is outside the yield surface, τ e > k , the material flows, and the rate of
deformation relates to the stress one-to-one.
Rigid, Perfectly Plastic Flow
The word “rigid” means that the model neglects elastic deformation. The
phrase “perfectly plastic” means that the model neglects strain hardening—that
is, the yield strength k remains constant as the material flows.
This model is perhaps the most successful model of plasticity, and is
widely used to analyze the forming processes of metals, and the load-carrying
December 7, 2014
rheological models 4
Plasticity
http://imechanica.org/node/17162
Z. Suo
capacity of metallic structures. The model also describes the yield-stress fluids,
such as toothpastes and greases. The model requires a single experimental input:
the yield strength. Flows under various boundary conditions have been analyzed
(Hill 1950, Tanner 2000).
Rigid, perfectly plastic flow in shear. We have studied rigid,
perfectly flow in a simple state of stress, such as shear. The stress-strain diagram
consists of two horizontal lines and infinitely many vertical lines. Because the
model neglects strain hardening, the stress-strain diagram is invariant with the
amount of strain. That is, the amount of flow does not affect the plastic state of
the material.
•
•
•
When the stress is within the yield strength, τ < k , the material is rigid,
and the strain does not change, γ = 0 . In the stress-strain diagram, this
part corresponds to a bi-directional, vertical line.
When the stress reaches the yield strength in the positive direction,
τ = +k , the strain can only increase, γ > 0 . The model of perfect plasticity
does not specify the amount of strain. In the stress-strain diagram, this
part corresponds to a horizontal line directing to the right.
When the stress reaches the yield strength in the negative direction,
τ = −k , the strain can only reduce, γ < 0 . In the stress-strain diagram,
this part corresponds to a horizontal line directing to the left.
τ
+k
γ
−k
Rigid, perfectly plastic flow in multiaxial state of stress. We have
generalized this stress-strain diagram in shear into a model for any state of stress.
Given a general state of stress s, we calculate the equivalent shearing stress,
τ e = sij sij / 2 . The material obeys the Mises yield condition, τ e = k , where k is
the yield strength under shear. The model neglects strain hardening, and the
yield strength k remains constant as the material flows. Consequently, the Mises
yield condition τ e = k corresponds to a level set in the stress space, known as the
yield surface.
December 7, 2014
rheological models 5
Plasticity
http://imechanica.org/node/17162
Z. Suo
D = λs
forbidden
rigid
yield surface
1
s s = k2
2 ij ij
stress space
The material obeys the Levy-Mises flow rule, D = λs , where λ is a scalar
measure of the flow, and depends on the state of stress as follows. The model
forbids the stress to go outside the yield surface. When the stress is inside the
yield surface, τ e < k , the material is rigid, and the rate of deformation vanishes,
λ = 0 . When the stress is on the yield surface, τ e = k , the material flows, λ > 0 ,
but the model does not specify the magnitude of λ .
The flow rule D = λs has a pictorial interpretation:
the rate of
deformation is a “vector” normal to the yield “surface”.
Rigid, Strain-Hardening Flow
Rigid, strain-hardening flow in monotonic shear. We have
studied this model for a material in a simple state of stress, such as shear. We
apply a monotonically increasing stress as a function of time, τ t , and measure
()
()
the strain as a function of time, γ t . The material is said to be rate-independent
if the stress-strain curve is unaffected by how the stress increases with time, so
long as τ t is an increasing function. For a rate-independent material, we plot
()
()
the experimental data as a stress-strain curve, τ = g γ .
When the stress is below the initial yield strength, the material is rigid,
and the stress increases along the vertical line, γ = 0 . We neglect elastic
deformation. We then increase the stress again, beyond the initial yield strength,
and the material flows. The material is said to strain-harden when the stressstrain curve monotonically increases.
For a strain-hardening, rate-independent material, define the tangent
modulus by Gt = dτ / dγ . The strain-hardening material has a monotonic stressstrain curve. We can regard the tangent modulus either as a function of stress or
December 7, 2014
rheological models 6
Plasticity
http://imechanica.org/node/17162
Z. Suo
as a function of strain. Here we will write the tangent modulus as a function of
stress, Gt τ . For a rate-independent material, write
()
τ = Gt τ γ .
()
()
τ =g γ
τ
()
Gt τ
1
γ
Rigid, strain-hardening flow in arbitrary history of shear.
Whereas the material is rate-independent, the response of the material depends
()
on the history of stress. Now let τ t be a history of stress not monotonic in
time. For example, we first increase the stress and cause the material to flow for
some time, and then reduce the stress. Upon unloading, the stress and strain do
not trace back along the curve; rather, the material becomes rigid, and the stress
decreases along a vertical line.
Upon reloading, the material is initially rigid, and then flows at the
previous point on the curve. That is, the previous flow has hardened the material
and increased the yield strength to a new level. The bent stress-strain curve in
monotonic shearing tells us how the yield strength k increases with flow.
Upon unloading to zero stress, we can shear the material in the opposite
direction. When the magnitude of the stress is small, the material is rigid, and
the stress goes done along vertical line. At some level of the shearing stress, the
material flows, and the stress-strain curve bends. As an idealization, we assume
that strain hardening is isotropic: the material yields at the current yield strength
established by prior deformation, and the bent stress-strain curve in the negative
direction is the same as that in the positive direction.
The above model of isotropic hardening achieves something
extraordinary: it uses the experimental record of a single history (i.e., the
monotonic loading) to describe all histories of loading, unloading, reloading, and
reverse loading. Given a history of the shearing stress, τ t , the model co-evolve
()
()
()
the history of strain γ t , and the history of the yield strength, k t .
December 7, 2014
rheological models 7
Plasticity
http://imechanica.org/node/17162
Z. Suo
τ
k
k
γ
Internal variable and plastic state. We describe a state of the
material using three parameters: the stress, the strain, and the yield strength. We
call the stress and the strain the external variables, which together let the
external world do work and change the state of the material. We call the yield
strength an internal variable.
If we only look at the monotonic loading, the yield strength and stress
coincide, so that the yield strength plays no distinct role. If we look at an
arbitrary history of loading and unloading, the stress and strength play distinct
roles, and the histories τ t and k t are different. We say that the internal
()
()
variable—the strength—characterizes the plastic state of the material (Prager
1949). As the material strain-hardens, the strength increases. The function k t
()
records the history of the plastic state.
We will use a single internal variable, the strength, to construct the
isotropic-hardening model. In more complex models of plasticity, we use more
than one internal variable to characterize the plastic state of the material.
Family of yield surfaces. We next generalize the model to an arbitrary
history of multiaxial stress, s t . Given a history of stress, we wish to construct a
()
model that co-evolves the history of multiaxial strain and the history of strength.
The model consists of there parts: the yield condition, the hardening rule, and
the flow rule.
The second-invariant model assumes that the material obeys the Mises
yield condition, τ e = k . For a strain-hardening material, the strength k increases
as the material flows. In the stress space, the yield condition corresponds to a
sequence of level sets, namely, a family of yield surfaces. The initial yield surface
December 7, 2014
rheological models 8
Plasticity
http://imechanica.org/node/17162
Z. Suo
is the yield surface of an as-received material, before we cause any flow. We will
also speak of subsequent yield surfaces, and the current yield surface. Each yield
surface corresponds to a distinct value of the strength.
1
s s = k t0
2 ij ij
( )
1
s s =k t
2 ij ij
()
Hardening rule. We next formulate the hardening rule to evolve the
strength k. Given the history of general state of stress, s t , we calculate the
()
history of the equivalent stress,
1
s t sij t ,
2 ij
as well as the rate of the equivalent stress, τe = dτ e t / dt .
()
() ()
τe t =
()
()
The model evolves strength k t incrementally.
We know the yield
()
strength at the current time, k t , we calculate
!
if τ e t < k t
# 0,
#
k = " 0,
if τ e t = k t and τe t < 0
#
# τe t , if τ e t = k t and τe t > 0
$
()
() ()
() ()
() ()
()
()
() ()
When the state of stress is inside the current yield surface, τ e t < k t , the
material is rigid, and the yield strength remains unchanged, k = 0 . When the
state of stress is on the current yield surface, τ e t = k t and but unloads,
() ()
τe t < 0 , the material is rigid, and the yield strength remains unchanged, k = 0 .
()
() ()
When the state of stress is on the current yield surface τ e t = k t and is loading,
τe t > 0 , the material flows strain hardens, increasing the strength at a rate
k = τ t .
()
e
()
December 7, 2014
rheological models 9
Plasticity
http://imechanica.org/node/17162
Z. Suo
Flow rule. The second-invariant model evolves the state of deformation
according to the Levy-Mises flow rule, D = λs . We calculate the scalar λ using
the experimentally measured stress-strain curve. The flow rule becomes
k
D=
s.
2kGt k
()
When the history of stress is of a special kind—the shearing stress as a monotonic
function of time—the flow rule reduces to τ = Gt τ γ .
()
The model is rate-independent. The prediction depends on the direction
of time, but not on the duration of time. That is, the prediction depends on the
sequence of ups and downs in the history of stress, s t , but is independent of
()
the durations between ups and downs.
The model achieves something extraordinary: it uses the experimental
record of a single history of a single type of stress (e.g., the stress-strain curve
under monotonic shearing) to predict all histories and all states of stress. Hill
(1950) credited this achievement to Schmidt (1932) and Odquist (1933).
Elastic-Plastic Flow
What does ε = ε e + ε p mean? A rod is of length L in the initial stressfree state, and is of length l in the current state under stress. The ratio of the two
lengths defines the stretch, λ = l / L .
L
lp
l
F
F
Starting from the current state, we remove the stress, so that the rod
unloads elastically and reaches an intermediate state of length l p . Both the
initial state and the intermediate state are stress-free, but the two states differ by
plastic deformation; define the plastic stretch by λ p = l p / L . The intermediate
December 7, 2014
rheological models 10
Plasticity
http://imechanica.org/node/17162
Z. Suo
state and the current state differ by an elastic deformation; define the elastic
stretch by λ e = l / l p .
Note an identity: λ = λ e λ p . Recall the definition of the natural strain,
ε = log λ . Similarly, write ε p = log λ p and ε e = log λ e . The expression λ = λ e λ p is
equivalent to ε = ε e + ε p .
Et
σ
Y
ε
E
Y
Elastoplasticity under uniaxial stress. We adopt the assumption
that the total strain is a sum of the elastic strain and plastic strain:
ε = εe +ε p .
The elastic strain relates to the stress as
σ
.
E
In the stress-strain diagram, we draw a straight line of slope E to represent elastic
loading and unloading. The projection of this line on the axis of strain is the
elastic strain. Because the stress is usually much smaller than Young’s modulus,
the straight line is nearly vertical, and the elastic stain is typically less than 1%.
The elastic strains in the two transverse directions are −νσ / E , where ν is
εe =
Poisson’s ratio.
If the current stress satisfies the yield condition, σ = Y , and the
increment of the stress causes plastic loading, d σ > 0 , the metal loads along the
December 7, 2014
rheological models 11
Plasticity
http://imechanica.org/node/17162
Z. Suo
bent curve. The increment of stress changes the yield strength by dY = dσ , and
changes the plastic strain by
"
%
1
1
dε p = $
− ' dσ .
$E Y
E '&
# t
( )
This expression gives the increment of strain in the axial direction. The plastic
strains in the two transverse directions are one half that in the axial direction.
Elastic, perfectly plastic flow in multiaxial state. We adopt the
assumption that the rate of deformation is the sum of elastic and plastic parts
(Hill 1959; Prager 1961):
D = De + D p .
The model calculates the rate of plastic strain as follows. Let σ ij be a state
of stress.
The mean stress is σ m = σ kk / 3 , and the deviatoric stress is
sij = σ ij − σ mδij .
Calculate the equivalent shearing stress, τ e = sij sij / 2 .
The
material obeys the Mises yield condition, τ e = k , where k is the yield strength
under shear. The model neglects strain hardening, and the yield strength k
remains constant as the material flows. The rate of plastic deformation obeys the
Levy-Mises flow rule, D p = λs . The model forbids the stress to go outside the
yield surface. When the stress is inside the yield surface, τ e < k , the material is
rigid, and the rate of deformation vanishes, λ = 0 . When the stress is on the yield
surface, τ e = k , the material flows, λ > 0 , but the model does not specify the
magnitude of λ .
The model calculates the rate of elastic deformation according to Hooke’s
law. For metals, dilatational strain and shearing strain are typically of
comparable magnitude. We will assume that elastic deformation is compressible,
but plastic deformation is incompressible. Thus, the rate of dilatation is linear in
the rate of the mean stress:
σ
Dkk = m ,
B
where B is the bulk modulus. The rate of deviatoric deformation is linear in the
rate of deviatoric stress:
∂s
1
Dije − Dkkδij = ij ,
3
2G
where G is the shear modulus. To preserve frame-indifference, we have used the
co-rotational rate:
∂s = ∂ins − Ws + sW .
December 7, 2014
rheological models 12
Plasticity
http://imechanica.org/node/17162
Z. Suo
Elastic, isotropic hardening flow in multiaxial state. Once again
we adopt the assumption that the rate of deformation is the sum of the elastic and
plastic parts:
D = De + D p .
We adopt the Mises yield condition τ e = k , and the same hardening rule as in the
model of rigid, isotropic hardening flow.
We also adopt the Levy-Mises flow rule, D p = λs , and assume that λ is a
( )
( )
function of the second invariant, λ τ e .
We fit the function λ τ e to the
experimental data under monotonic shear. In the shearing flow, the model
reduces to
τ
τ
, γe = .
γ = γe + γ p , γ =
G
Gt τ
()
Consequently, the rate of plastic deformation relates to the rate of stress as
"
%
1
1'
p
$

γ =
− τ .
$ Gt τ G '
#
&
When the material is subject to a general history of multiaxial stress, the flow rule
becomes:
"
%
1
1 ' k
p
$
D =
−
s.
$ G k G ' 2k
# t
&
We specify the rate of elastic deformation according to Hooke’s law:
σ
Dkk = m ,
B
∂s
1
Dije − Dkkδij = ij .
3
2G
To preserve frame-indifference, we adopt the co-rotational rate:
∂s = ∂ins − Ws + sW .
()
()
Elastic-plastic deformation is history-dependent. One can pull
and twist a thin-walled tube to achieve various histories of homogeneous
deformation. This experimental setup has played a significance role in the
development of plasticity (Hill 1950).
The elastic, plastic response is history-dependent. Describe a history of
stress by the tensile stress and shearing stress as functions of time, σ t and
()
()
τ t . Describe a history of deformation by the tensile strain and shearing strain
()
()
stress is a curve on the (σ ,τ ) plane, and a history of strain is a curve on the (ε ,γ )
as functions of time, ε t and γ t . For a rate-independent material, a history of
December 7, 2014
rheological models 13
Plasticity
http://imechanica.org/node/17162
Z. Suo
plane. Each history is fully specified by a curve and its direction. Each point on
the curve corresponds a state of stress or strain. How fast the material goes from
one state to another does not matter.
Given a history of stress, the elastic, perfectly plastic model cannot predict
a unique history of strain. Because once the material reaches the yield condition,
the amount of deformation is arbitrary. Given a history of strain, however, the
elastic, perfectly plastic model predicts a unique history of stress.
τ
γ
B
B
C
O
A
C2
C1
ε
O
A
σ
Consider two histories strain. In history OAC, we pull the tube to the
state just below the yield condition, hold the tensile strain, and then twist. In
history OBC, we twist the tube to the state just below the yield condition, hold the
shearing strain, and then pull. The two histories reach the same state of strain at
the end, C.
The elastic, perfectly plastic model predicts the two corresponding
histories of stress. The two histories will end at two different states of stress, C1
and C2.
Let’s calculate. The Mises yield condition is
1 2 2
σ + τ = k2 .
3
The yield locus is an ellipse in the stress plane. The stress-strain relation is
σ 2
τ
ε = + λσ , γ = + 2λτ .
E 3
G
Here we neglect the rate of rotation, and replace the co-rotational rate with the
in-frame rate. When the state of stress is inside the yield locus, the material is
elastic and λ = 0 . When the state of stress moves on the yield surface, the
material flows, and λ > 0 . Given a history of strains, ε t and γ t , the above
()
()
three equations determine the histories of λ (t ) , σ (t ) and τ (t ) .
To simplify algebra, assume the material in incompressible, so that
E = 3G . Eliminating λ by combining the two equations in the flow rule, we
obtain that
December 7, 2014
rheological models 14
Plasticity
http://imechanica.org/node/17162
Z. Suo
3Gε − σ σ
=
Gγ − τ τ
Along the segment AC, the tensile strain is fixed, ε , and the state of stress moves
along the yield locus, σ 2 / 3 + τ 2 = k 2 .
Integrating
over
the
k 2τ
= Gγ
k2 − τ 2
interval 0 < γ < k / G ,
we
get
the
state
C 1:
σ = 0.76 3k, τ = 0.65k . We can similarly get state C2: σ = 0.65 3k, τ = 0.76k .
Ratcheting plastic deformation. Describe Huang, Suo, Ma (2002).
Viscoelastic Flow
In a separate course, we study viscoelasticity in infinitesimal deformation
(Suo 2008). Many powerful ideas in linear viscoelasticity become invalid when
materials are nonlinear. Here we will focus on one idea that is valid for large
deformation: representing rheological model using springs and dashpots (Reiner
1945).
For example, consider the Maxwell model, represented by a spring and a
dashpot in series. We interpret this picture as
D = De + Dv ,
where De represents the rate of elastic deformation, and D v represents the rate
of viscous deformation. We calculate the rate of viscous deformation using the
second-invariant model:
s
Dv =
.
2η
We calculate the rate of elastic deformation using Hooke’s law:
σ
Dkk = m
B
∂s
1
Dije − Dkkδij = ij .
3
2G
To preserver the frame-indifference, we adopt the co-rotational rate.
∂s = ∂ins − Ws + sW .
A commonly used model for metals and ceramics at elevated
temperatures is to model viscosity by power law, and model elasticity using
Hooke’s law. The rate of deformation is often negligible, so that the co-rotational
rate is replaced with the in-frame rate.
We can, of course, adopt other arrangements of springs and dashpots,
other models of viscosity and elasticity, and other frame-indifferent rates. We get
December 7, 2014
rheological models 15
Plasticity
http://imechanica.org/node/17162
Z. Suo
many viscoelastic models (Tanner 2000, Irgens 2014). Sorting what models
predict what phenomena is not easy, and is still an ongoing topic of research.
Inhomogeneous Deformation
When a body of the material undergoes inhomogeneous deformation, we
regard the body as a sum of many small pieces. Each small piece undergoes
homogeneous deformation, and obeys the rheological model specified above.
Different pieces in the body communicate through the compatibility of
deformation and the balance of forces. We use the Eulerian approach.
Time derivative of a function of material particle. At time t, a
material particle X moves to position x (X , t ) . The velocity of the material
particle is
v=
( ).
∂x X,t
∂t
Here, the independent variables are the time and the coordinates of the material
Let G (X , t ) be a function of material particle and time. For example, G
can be the temperature of material particle X at time t. The rate of change in
temperature of the material particle is
∂G (X , t )
.
∂t
This rate is known as the material time derivative.
We can calculate the material time derivative by an alternative approach.
Change the variable from X to x by using the function X (x , t ), and write
g(x, t ) = G(X, t )
Using chain rule, we obtain that
∂G (X , t ) ∂g(x , t ) ∂g(x , t ) ∂x i (X , t )
.
=
+
∂t
∂t
∂x i
∂t
In particular, we replace the in-frame rate of stress with the material
derivative, so that the correlational rate becomes
∂s x,t
∂s x,t
∂s =
+ vi
− Ws + sW .
∂t
∂xi
( )
( )
Compatibility of deformation. Let x be the coordinate of a place in
space, t be the time, and vi x,t be the velocity of a small piece of fluid at the
( )
place x and time t. The compatibility of deformation relates the rate of
deformation to the field of velocity:
December 7, 2014
rheological models 16
Plasticity
http://imechanica.org/node/17162
Z. Suo
1 " ∂v ∂v %
Dij = $$ i + j '' .
2 # ∂x j ∂xi &
This relation between the rate of deformation and velocity holds for deformation
of any magnitude.
( )
Balance of forces. Let bi x,t be the body force per unit volume, and
ρ be the mass per unit volume. For an incompressible fluid, ρ is constant.
Inside the body, the balance of forces relates the stress to the body force and the
inertial force:
" ∂v
∂σ ij
∂v %
+ bi = ρ $$ i + v j i '' .
∂x j
∂x j &
# ∂t
The right-hand side is the inertial force per unit volume.
Let ni x,t be the unit vector normal to a small part of the surface of the
body.
( )
Let t ( x,t ) be
i
the traction, i.e., the force per unit area acting on the
surface. On the surface of the body, the balance of forces relates the stress to the
traction:
σ ij n j = ti .
Boundary conditions. We divide the surface of the body into two
parts. On one part of the surface, Sv , we prescribe the velocity of the fluid. On
the other part of the surface, St , we prescribe the traction. These boundary
conditions, along with the equations listed above, constitute a boundary-value
problem that governs the inhomogeneous deformation in the body.
References
M.E. Gurtin, E. Fried and L. Anand, The Mechanics and Thermodynamics of
Continua. Cambridge University Press, 2010.
R. Hill, Mathematical Theory of Plasticity, Oxford University Press, 1950.
R. Hill, Some basic principles in the mechanics of solids without a natural time.
Journal of the Mechanics and Physics of Solids 7, 209-225 (1959).
M. Huang, Z. Suo, Q. Ma. Plastic ratcheting induced cracks in thin film
structures. Journal of the Mechanics and Physics of Solids 50, 1079-1098
(2002).
F. Irgens, Rheology and Non-Newtonian Fluids. Springer, 2014.
S. Nemat-Nasser, Plasticity. Cambridge University Press, 2004.
W. Prager, Recent developments in the mathematical theory of plasticity.
Journal of Applied Physics 20, 235-241 (1949).
December 7, 2014
rheological models 17
Plasticity
http://imechanica.org/node/17162
Z. Suo
W. Prager, An elementary discussion of definitions of stress rate. Quarterly
Applied Mathematics 403-407 (1961).
M. Reiner, A classification of rheological properties. Journal of Scientific
Instruments 22, 127-129 (1945).
G.G. Stokes, On the theories of the internal friction in motion, and of the
equilibrium and motion of elastic solids. Transactions of the Cambridge
Philosophical Society 8, 287-341 (1845).
Z. Suo, Viscoelasticity, http://imechanica.org/node/482 (2008).
Z. Suo, Advanced Elasticity, http://imechanica.org/node/725 (2013).
Z. Suo, Plasticity, http://imechanica.org/node/17162 (2014).
R.I. Tanner, Engineering Rheology, 2nd edition. Oxford University Press, 2000.
H. Xiao, O.T. Bruhns, A. Meyers, Elastoplasticity beyond small deformations.
Acta Mechanics 182, 31-111 (2006).
December 7, 2014
rheological models 18
Download